The alternative hypotheses tested are the cohesion hypothesis, which predicts a decrease in criminal violence in times of increased security-related stress, and the legitimation-habituation hypothesis, which predicts an increase in criminal violence in such periods. The study uses monthly data collected in Israel over 15 years. Security-related casualties and incidents are the security-related stressors; homicide and robbery are the measures of criminal violence; and economic and sociodemographic indicators are the control variables. The analysis uses a multiple regression model with autoregressive errors. The reported number of casualties had a significant marginal positive effect on homicide, thus supporting the legitimation-habituation hypothesis. Such an effect could not be detected for the robberies series. Inflation increments had a strong positive effect on both series, and rates of unemployment had a significant marginal positive effect on robbery but not on homicide. The effects of the various stressors are not contemporaneous but extend to several months ahead. Some theoretical aspects of the findings and their implications for Israeli society are discussed. 3 tables, 46 references. (Author abstract modified)